A number of philosophers have appealed to “second-order desires” (desires about desires) in order to explain basic moral notions. Harry Frankfurt has suggested that freedom of the will might be identified with the ability to satisfy a certain sort of second-order desire and Richard Jeffrey has made a similar proposal. Gary Watson has objected that the resulting account of freedom of the will does not work and that the appeals made to second-order desires should be replaced with references to what an agent values.

Watson's view can be reconciled with Frankfurt's if valuing is identified with a kind of second-order desire. David Lewis has offered such an identification and David Copp has recently suggested that we might identify a person's values with preferences for which the person has a certain second-order preference.

However, these appeals to second-order desires in order to explain freedom of the will, valuing, or values are all unsuccessful.

In arguing for this negative conclusion, I defend a number of subsidiary points: (1) Positive intentions are reflexive or self-referential. (2) Intrinsic desires are not just noninstrumental desires. (3) To value something is in part to want it. (4) There is sometimes a difference between believing that something is good and valuing it. (5) There is a difference between valuing something and having it as one of your values.

Frankfurt on freedom of the will

Frankfurt motivates his view by criticizing Strawson's account of the concept of a person.